Cargando…
De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates
When sparsely sampled data are used to accelerate magnetic resonance imaging (MRI), conventional reconstruction approaches produce significant artifacts that obscure the content of the image. To remove aliasing artifacts, we propose an advanced convolutional neural network (CNN) called fully dense a...
Autores principales: | , , , , , |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
MDPI
2022
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9854709/ https://www.ncbi.nlm.nih.gov/pubmed/36671594 http://dx.doi.org/10.3390/bioengineering10010022 |
_version_ | 1784873189084495872 |
---|---|
author | Hossain, Md. Biddut Kwon, Ki-Chul Imtiaz, Shariar Md Nam, Oh-Seung Jeon, Seok-Hee Kim, Nam |
author_facet | Hossain, Md. Biddut Kwon, Ki-Chul Imtiaz, Shariar Md Nam, Oh-Seung Jeon, Seok-Hee Kim, Nam |
author_sort | Hossain, Md. Biddut |
collection | PubMed |
description | When sparsely sampled data are used to accelerate magnetic resonance imaging (MRI), conventional reconstruction approaches produce significant artifacts that obscure the content of the image. To remove aliasing artifacts, we propose an advanced convolutional neural network (CNN) called fully dense attention CNN (FDA-CNN). We updated the Unet model with the fully dense connectivity and attention mechanism for MRI reconstruction. The main benefit of FDA-CNN is that an attention gate in each decoder layer increases the learning process by focusing on the relevant image features and provides a better generalization of the network by reducing irrelevant activations. Moreover, densely interconnected convolutional layers reuse the feature maps and prevent the vanishing gradient problem. Additionally, we also implement a new, proficient under-sampling pattern in the phase direction that takes low and high frequencies from the k-space both randomly and non-randomly. The performance of FDA-CNN was evaluated quantitatively and qualitatively with three different sub-sampling masks and datasets. Compared with five current deep learning-based and two compressed sensing MRI reconstruction techniques, the proposed method performed better as it reconstructed smoother and brighter images. Furthermore, FDA-CNN improved the mean PSNR by 2 dB, SSIM by 0.35, and VIFP by 0.37 compared with Unet for the acceleration factor of 5. |
format | Online Article Text |
id | pubmed-9854709 |
institution | National Center for Biotechnology Information |
language | English |
publishDate | 2022 |
publisher | MDPI |
record_format | MEDLINE/PubMed |
spelling | pubmed-98547092023-01-21 De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates Hossain, Md. Biddut Kwon, Ki-Chul Imtiaz, Shariar Md Nam, Oh-Seung Jeon, Seok-Hee Kim, Nam Bioengineering (Basel) Article When sparsely sampled data are used to accelerate magnetic resonance imaging (MRI), conventional reconstruction approaches produce significant artifacts that obscure the content of the image. To remove aliasing artifacts, we propose an advanced convolutional neural network (CNN) called fully dense attention CNN (FDA-CNN). We updated the Unet model with the fully dense connectivity and attention mechanism for MRI reconstruction. The main benefit of FDA-CNN is that an attention gate in each decoder layer increases the learning process by focusing on the relevant image features and provides a better generalization of the network by reducing irrelevant activations. Moreover, densely interconnected convolutional layers reuse the feature maps and prevent the vanishing gradient problem. Additionally, we also implement a new, proficient under-sampling pattern in the phase direction that takes low and high frequencies from the k-space both randomly and non-randomly. The performance of FDA-CNN was evaluated quantitatively and qualitatively with three different sub-sampling masks and datasets. Compared with five current deep learning-based and two compressed sensing MRI reconstruction techniques, the proposed method performed better as it reconstructed smoother and brighter images. Furthermore, FDA-CNN improved the mean PSNR by 2 dB, SSIM by 0.35, and VIFP by 0.37 compared with Unet for the acceleration factor of 5. MDPI 2022-12-22 /pmc/articles/PMC9854709/ /pubmed/36671594 http://dx.doi.org/10.3390/bioengineering10010022 Text en © 2022 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). |
spellingShingle | Article Hossain, Md. Biddut Kwon, Ki-Chul Imtiaz, Shariar Md Nam, Oh-Seung Jeon, Seok-Hee Kim, Nam De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates |
title | De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates |
title_full | De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates |
title_fullStr | De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates |
title_full_unstemmed | De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates |
title_short | De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates |
title_sort | de-aliasing and accelerated sparse magnetic resonance image reconstruction using fully dense cnn with attention gates |
topic | Article |
url | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9854709/ https://www.ncbi.nlm.nih.gov/pubmed/36671594 http://dx.doi.org/10.3390/bioengineering10010022 |
work_keys_str_mv | AT hossainmdbiddut dealiasingandacceleratedsparsemagneticresonanceimagereconstructionusingfullydensecnnwithattentiongates AT kwonkichul dealiasingandacceleratedsparsemagneticresonanceimagereconstructionusingfullydensecnnwithattentiongates AT imtiazshariarmd dealiasingandacceleratedsparsemagneticresonanceimagereconstructionusingfullydensecnnwithattentiongates AT namohseung dealiasingandacceleratedsparsemagneticresonanceimagereconstructionusingfullydensecnnwithattentiongates AT jeonseokhee dealiasingandacceleratedsparsemagneticresonanceimagereconstructionusingfullydensecnnwithattentiongates AT kimnam dealiasingandacceleratedsparsemagneticresonanceimagereconstructionusingfullydensecnnwithattentiongates |